[rancid] Dell 3548P Switches with Rancid

brain conflict brconflict at gmail.com
Wed Jul 18 15:23:15 UTC 2012


Got a strange issue, that I'm sure others have come across:

With a Dell 3548P switch, code rev: 2.0.0.29, "stacked", Rancid (v.
2.3.2) is having an issue backing them up. The error is that the
switches have not been successfully contacted in x hours.

What I found in debugging and test (Switch IP is 10.2.2.1, btw):
dlogin -t 90 -c"show version" 10.2.2.1 attempts to connect to the
switch, logs in successfully, but doesn't do anything beyond that. I
looked at a tcpdump from the server RANCID runs on and found that,
(using telnet for non-encrypted viewing), the switch responds to the
successful login with its prompt, "switch#". The server does nothing
beyond that, and soon Errors on a "TIMEOUT reached". I'm concerned
that the script doesn't actually "see" the correct prompt, or that
there's something of a hack that needs to be implemented beyond the
latest Dell modules from ftp.shrubbery.net/pub/rancid/contrib.

Any ideas? I'm about to test this with a single 3548P, not stacked,
then on an older version of code. The 6248 switches operate fine here,
even stacked.

-brconflict


More information about the Rancid-discuss mailing list